WO2016183654A1

WO2016183654A1 - System and method for identifying the characteristics of an electric machine

Info

Publication number: WO2016183654A1
Application number: PCT/BR2016/050106
Authority: WO
Inventors: Mauricio RIGONI; Nelson Jhoe Batistela; Nelson SADOWSKI; Leonardo Augusto FELER; Conrado Simões Pereira GAMEIRO; Helton Fernando Dos SANTOS; Ricardo De Araujo Elias; Patrick Kuo-Peng; João Pedro Assumpção BASTOS; Luciano Mendes De FREITAS
Original assignee: Tractebel Energia S.A.; Universidade Federal De Santa Catarina; Itá Energética S.A.
Priority date: 2015-05-19
Filing date: 2016-05-16
Publication date: 2016-11-24
Also published as: BR102015011438B1; BR102015011438A2

Abstract

The present invention provides a system and method for identifying the characteristics of an electric machine (3), comprising at least one non-invasive measuring device (1, 11, ..., 1N, 101) and one analysis device (4), wherein the at least one measuring device (1, 11, ..., 1N, 101) acquires at least one time derivative value with at least one frequency spectrum component of the waveform of an electromagnetic quantity (50, 51, 52, 53, ..., 5N, 501) generated by the electric machine.

Description

Report of the Invention Patent for "SYSTEM AND METHOD FOR IDENTIFYING CHARACTERISTICS OF AN ELECTRIC MACHINE".

[001] The present invention relates to a system and method for identifying established or incipient electrical and mechanical characteristics in electrical machines by analyzing disturbances in the external magnetic field measured with sensors using non-external external measuring tools. invasive.

Description of the prior art

[002] Electrical machines are monitored by various types of systems to identify faults, faults, operating condition and operation, with a view to maintaining their operability and useful life. Many monitoring systems only operate when the machine has already had a fault that makes it impossible to operate. These commonly used systems are generally not able to detect incipient failures, especially those that do not cause them to stop. Other monitoring systems serve to detect incipient faults such as mechanical vibration, air gap displacement, short circuits, among others. Most of these types of monitoring use invasive sensors or monitor physical quantities other than magnetic field / induction to detect machine anomalies. There are methodologies and magnetic field sensors, or commercial magnetic induction sensors, or for application in academic studies, which monitor field or magnetic induction, detecting machine anomalies.

[003] In the past, solutions were sought for problems with the installation of high power generators, which had as their main defect the overheating due to short circuits. To solve this problem, we sought to detect the short circuits between turns by evaluating the magnetic fields in the air gap, around the core and stator windings, always through sensors invasive to the machine. The difficulty came in having to disassemble the electric machine so that the sensor could be inserted somewhere inside the housing, such as the stator or a tooth or slot in the electric machine. This makes it difficult to install and replace these sensors, which requires a complete shutdown of the machine, involving large operating costs and investments.

The practice of associated magnetic field analysis for fault detection in induction motors such as related to broken bars and stator turns short circuits has begun to be explored by the ease of both deploying a external sensor in machines already installed as for the functional characteristics of this type of electric machine (because of the slip phenomenon, unlike synchronous machines, where induction motors are asynchronous machines, it is easier to detect faults in this motor type by external magnetic field analysis). In this type of external field monitoring, there is no need to disassemble the electrical machine for a sensor to be inserted.

An example of a prior art invasive sensor is known from patent document EP1418655 which discloses an invasive method using a sensor called a flow probe which is inserted into the electrical machine installed on the generator stator to measure the magnetic field and variations of field values.

WO2013136098 discloses an invasive method for detecting damage in alternating rotary machines by differential magnetic field measurement using two measuring coils mounted on the stator teeth, ie within the machine housing. [007] US2009243647 discloses a method and system for identifying defects in an electrical machine based on magnetic field monitoring through sensors disposed outside the machine that evaluate the generated magnetic field. However, this document only mentions measuring the values of the magnetic field itself, to determine if the electric machine is missing and not other parameters related to the magnetic field.

WO2012097825 discloses a circuit and a method of detecting faults in a wind turbine. The fault detection circuit comprises a magnetometer in the form of a Hall effect sensor coupled between a power converter and a power converter ground element and is configured to measure a ground current from the power converter to obtain a ground current; and a comparator configured to determine the presence of a fault based on the actual ground current. However, this document does not mention the measurement of other parameters related to the ground current, rather than the ground current itself.

Objectives of the invention

It is an object of the present invention to provide a non-invasive method and system for detecting faults, malfunctions and defects in electrical machines.

It is also an object of the present invention to identify different types of electrical machine fault by non-invasive measurements.

It is a further object of the present invention to provide a simpler signal capture system for determining the health of the electric machine.

Brief Description of the Invention The objects of the invention are achieved by means of a system for identifying characteristics of an electrical machine, comprising at least one non-invasive meter and an analysis device, wherein the at least one meter identifies time derivative values of a quantity. generated by the electric machine and sends these time derivative values of the electromagnetic quantity to the analysis device, and the device interprets the time derivative values of the electromagnetic quantity identified by at least one noninvasive meter and determines at least one characteristic of the electric machine.

The time derivative values of electromagnetic magnitude include at least one component of the waveform frequency spectrum of one of the magnetic field external to the electric machine, the magnetic flux external to the electric machine, the magnetic induction external to the electric machine. and the grounding current of the electrical machine, where each meter measures at least one different component of the frequency spectrum of the electrical machine.

The analysis device compares the identified time derivative values of the electromagnetic quantity with pre-established time derivative values of this electromagnetic quantity and determines at least one characteristic of the electric machine based on the result of the comparison.

The time derivative values of the identified electromagnetic quantity correspond to the n-order temporal derivative of said electromagnetic quantity, where n ranges from 1 to infinity.

In addition, the at least one noninvasive meter is at least one of: a one-way, three-dimensional point meter, and each at least one noninvasive meter comprises at least one of: [0017] an analog sensor identifying the time derivative values of the electromagnetic quantity and

[0018] a sensor comprising a means of calculation, wherein the sensor measures scaling or vectorically the values of the electromagnetic magnitude, and the means of calculation applies derivative operators to the values measured by the sensor, and identifies the time derivative values of the electromagnetic magnitude.

Each at least one non-invasive meter is arranged in a position between a location in contact with the machine casing and a distance away from the electric machine of up to 20 times the diameter of the electric machine.

Alternatively each at least one noninvasive meter is arranged in a position surrounding the grounding cable of the electrical machine.

[0021] The identified electrical machine characteristic is at least one of: non-standard machine operating diagnostic characteristics, lifetime, incipient fault onset characteristics, fault state, or machine operating characteristics.

[0022] The objects of the invention are further achieved by a method for identifying features of an electric machine, wherein the method comprises the following steps:

- identify time derivative values of at least one electromagnetic quantity generated by the electric machine in at least one position external to the electric machine;

- comparing the identified values of the time derivative of at least one electromagnetic quantity with the pre-established values of the time derivative of this electromagnetic quantity generated by the electric machine; and

[0025] - determine at least one machine feature based on the result of the comparison.

The step of identifying the time derivative value of electromagnetic magnitude is performed by at least one noninvasive meter from:

[0027] an analog sensor for identifying time derivative values of an electromagnetic quantity and

[0028] a sensor comprising a means of calculation, wherein the sensor measures the values of the electromagnetic magnitude, and the means of calculation applies derivative operators to the values measured by the sensor, and numerically identifies the time derivative values of that electromagnetic magnitude.

Alternatively each meter identifies a different component of the frequency spectrum of the electrical machine, or derivatives of the measured electrical quantities of that machine.

The at least one position external to the electric machine where the step of identifying the time derivative values of at least one electromagnetic quantity is performed is comprised between a location in contact with the machine housing and a distance away from the electric machine of maximum 20 times the diameter of the electric machine.

The step of identifying time derivative values of at least one electromagnetic quantity is performed around the grounding cable of the electrical machine at a location between the grounding cable surface and a distance of up to 20 times the diameter of the electrical cable. electric machine.

Furthermore, the step of comparing identified time derivative values of an electromagnetic quantity generated by the electric machine with the pre-established time derivative values of such an electromagnetic quantity further comprises the substeps of: [0033] - filtering the measured signal;

[0034] - amplifying the signal;

Passing the signal through an anti-aliasing filter; and

[0036] - sample the signal.

Preferably and additionally the method comprises a step of storing the identified time derivative values of the electromagnetic quantity generated by the electric machine for use as pre-established time derivative values of the electromagnetic quantity.

The preset time derivative values of the electromagnetic quantity correspond to the operating measurement values of the sound electric machine, operating history values of the electric machine or expected theoretical values.

The method may be carried out by the system of the invention, object of the present invention.

Brief Description of the Drawing

The present invention will be described in more detail below. The figures show:

Figure 1 is a system diagram in a preferred embodiment of the invention whereby values, preferably of waveforms, of time derivative of an electromagnetic quantity generated by the electric machine are identified;

Figure 2 is a system diagram in another preferred embodiment whereby waveforms of the time derivative of the grounding current of the electrical machine are obtained.

[0043] Figure 3 is a block diagram of the control means used in the method according to the invention.

[0044] Figure 4 - is a block diagram of the control steps performed in a preferred embodiment of the method according to invention.

[0045] Figure 5 is a block diagram of the control steps performed in another preferred embodiment of the method according to the invention.

Figure 6 is an example of the electromagnetic signature identified by the system and method of the present invention of a synchronous generator operating at 60 Hz.

[0047] Figure 7 is a comparative example of electromagnetic signatures identified by the system and method of the present invention of two synchronous generators operating at 60 Hz.

Figure 8 is a comparative example of the electromagnetic signature of two 60 Hz synchronous generators obtained by the system and method according to one embodiment of the present invention, wherein the time derivative of an identified electromagnetic quantity is the derivative. of the electrical current of the grounding cable of the generators.

[0049] Figure 9 is an example of identifying the amplitude variation of subharmonic components as a function of the severity of a shorted field winding of a synchronous generator.

Detailed Description of the Drawing

In Figure 1, a system diagram for identifying characteristics of an electrical machine in a preferred embodiment is shown, whereby values, preferably of waveforms, of time derivative of an electromagnetic quantity 50, 51, are identified. 52, 53, ... 5N generated by the electric machine 3, and to which the method for identifying characteristics of an electric machine 3 according to the invention may be applied. Frequency spectrum components of the electrical machine comprise fundamentals, harmonics, subharmonics, and frequency interharmonics. mechanical and electrical Electromagnetic quantities are identified and analyzed in dB.

The structure and operation of the system will be described herein for ease of understanding as well as the method according to the invention, making it clear that the application of the method is not limited to the system described herein.

The system of figure 1 comprises one or more non-invasive meters 1, 1 1, 12, ..., 1 N externally positioned electric machine 3 in at least one position 2, 22, ..., 2N in the direction with convenient orientation of the meters or sensors to the direction of the magnetic field. This gauge position may vary between a location in contact with the machine housing and a distance away from the machine. This position 2, 22, ..., 2N may vary radially with respect to machine 3 from a point abutting the machine housing to a distance 20 times the diameter of machine 3.

The electromagnetic quantities measured and evaluated by the meters or sensors of the system of the present invention are: the fundamental frequency, fundamental harmonic, subharmonic and interharmonic components of the magnetic field, magnetic flux, induction waveforms. and grounding current in the grounding cable, or the fundamental and components of the mechanical frequency related to the generator rotation. The values used for the assessment and diagnosis of the operation of the electrical machine will therefore be the time derivative waveforms of these quantities and the frequency spectrum waveform components of the time derivatives of these quantities. The meters can directly measure these derivative values or waveforms, or they can measure the harmonic, subharmonic, and interharmonic components in question and perform the processing. data to obtain the values of the derivatives of these components.

The 1, 1 1, ..., 1 N gauges may be unidirectional, three-dimensional, or spot depending on the type of sensor used. In a preferred embodiment, the meter 1, 11, ..., 1 N is an analog sensor that directly identifies the values of the time derivative, for example of the first order, of an electromagnetic magnitude 50, 51, 52, 53. ..5N.

In another embodiment of the invention, meter 1, 11, ..., 1 N comprises a sensor and calculation means. The sensor measures the waveform, in scalar values or vector values of the electromagnetic quantity (1 D, 2D or 3D, by sensor association and by the very nature of the sensor), and the calculation means apply derivative operators to the measured values and perform data processing steps to identify the time derivative values of the electromagnetic quantity 50, 51, 52, 53, ... 5N.

In a preferred embodiment of the invention, the time derivative values of the electromagnetic magnitude 50, 51, 52, 53, ... 5N are identified by an induction sensor, such as a search-coil voltage. ). Other sensors such as magnetic field, magnetic flux or magnetic induction sensors, magnetic field / induction transducer elements, magnetic induction sensor, or air core, dielectric or ferromagnetic probe coil, among others , can be employed in the system meters.

In one embodiment of the invention, the time derivative values of the electromagnetic quantity 50, 51, 52, 53, ... 5N identified by meter (s) 1, 1 1, ..., 1 N correspond to to the n-order derivative of the external magnetic field generated by the electric machine 3. In general, the time-derivative values of the electro- 50, 51, 52, 53, ... 5N identified by meter (s) 1, 1 1, ..., 1 N may correspond to the n-order temporal derivative of the evaluated electromagnetic quantity, for example of the flux or magnetic induction generated by the electric machine 3 and measured externally to it. The value of n (order of the derivative) is from 1 to infinity for these cases, preferably being 1.

The position 2, 22, ..., 2N of each meter 1, 1 1, ..., 1 N with respect to the machine is identified as a function of distance in any radial orientation with respect to the opposite end of the housing of the and specifically the stator housing, as shown in the horizontal axis below the machine shown in Figure 1. These meters may be located at a position at a distance from electrical machine 3 that varies from a point in contact with the machine housing up to a radial distance of 30 times the diameter of the electric machine.

Each electric machine has an electromagnetic signature, which varies mainly due to the different types of machines, the different specifications, the different construction nature, the possible incipient failures, the state of life or the mode of operation of these machines. machines. Such electromagnetic signature is to be understood within the scope of this invention as the frequency spectrum of the time derivatives of the 50, 51, 52, 53, ... 5N waveforms identified by at least one meter 1, 1 1, .. ., 1 N. Examples of electromagnetic signatures can be seen in figures 6 to 9.

In a preferred embodiment of the invention, the electromagnetic signature of the electric machine 3 comprises one or more of the sub-synchronous, harmonic and interharmonic components of the electric fundamental frequency, including the null frequency component. These frequencies are also related to the frequency mechanical fundamental, that is, the mechanical fundamental and its harmonics.

Whenever a problem occurs in the electric machine 3, the electromagnetic signature varies compared to the sound machine signature. The present invention can identify this variation in the electromagnetic signature of machine 3 and relate this variation to a characteristic of the altered electric machine. The characteristic of the electric machine will be better described later.

In order to identify machine characteristics 3, the system according to the invention utilizes one or more of the meters 1, 1 1, ..., 1 N which identify the time derivative values of the electromagnetic quantity. , 51, 52, 53, ... 5N generated by the electric machine 3. After being identified, either directly by the meter sensor or by means of calculations coupled to the sensor, the time derivative values of an electromagnetic quantity 50, 51, 52, 53, ... 5N are sent to the analysis device 4. In a preferred embodiment, the use of only one sensor may be sufficient to achieve the analysis objectives 4.

Analyzer 4 interprets the time derivative values of an electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) generated by the machine, and / or of frequency spectrum components of waveforms. time derivative wave of these quantities to determine the characteristics of the electric machine 3. In a preferred embodiment of the invention the analysis device 4 treats and processes the values or waveform of the time derivative of the electromagnetic quantity 50, 51, 52, 53, ... 5N generated by the electrical machine 3 and / or of waveform frequency spectrum components of the time derivatives of these quantities, through filters 41 and instrumentation amplifiers 42. Shortly thereafter, the signal goes through an anti-aliasing filter 43 to avoid spectral overlay problems. Then the signal is sent to a sampler circuit 44 so that the signal is sampled and discretized. This sampler circuit 44 may be an analog / digital converter. A computational medium 45, such as a computer with a suitable processor and software, reads the signal from the sampler circuit 44, and compares this signal with the pre-set time derivative values of that machine-generated electromagnetic quantity. 3. In another embodiment of the invention, the computational medium 45 reads the signal from the sampler circuit 44, makes the frequency spectral decomposition of the signal, generating the magnetic signature, and compares that signature and / or values of spectrum components. frequency values with various corresponding frequency component values of the time derivative values of an electromagnetic quantity 50, 51, 52, 53, ... 5N external to machine 3, obtained previously or pre-established. The predefined time derivative values of an electromagnetic quantity 50, 51, 52, 53, ... 5N generated by the electric machine 3 and / or the waveform frequency spectrum components of the time derivatives of these quantities will be further described. forward.

In a preferred embodiment of the invention the computational medium 45 reads the amplitude values of the sub-synchronous, harmonic and interharmonic components of the electromagnetic magnitude of the signal generated by the sampler circuit 44 and compares them to the amplitude values of the pre-values. -established time derivatives of an electromagnetic quantity 50, 51, 52, 53, ... 5N generated by the electric machine 3.

After comparison, the computational medium 45 is able to identify the characteristics of the electric machine 3 and make them available. for viewing on the analysis device 4 through a display, or by another way of displaying analysis results, such as, for example, a listing of values. In another embodiment of the invention the computing medium 45 makes the data available to a supervisory system 6 whether or not included in the analyzing device 4. The analyzing device 4 ensures the proper signal-to-noise ratio necessary for the efficient acquisition of information for the computing medium 45 and for the supervisory system 6.

The supervisory system 6 reads the information transmitted to it and, depending on the characteristic of the electrical machine determined, if it is a fault characteristic of the electrical machine, triggers an alarm, which may be preferably audible or visual, or in report form. , or by another warning mode, advising a user of the fault characteristic of the electric machine.

The preset time derivative values of an electromagnetic quantity generated by the electric machine 3 can be obtained by a first measurement when the machine is in healthy condition, in order to obtain reference values. For example, when the machine is healthy, this measurement can be performed at a frequency of time that can be daily, weekly, monthly or yearly, or as per the needs / specifications of the user / expert. The preset values of time derivative of an electromagnetic quantity generated by electric machine 3 can also be obtained through a history of stored values while the machine is healthy. In addition, the predefined time derivative values of an electromagnetic quantity generated by the electric machine 3 may also consist of expected theoretical values, which may be previously calculated or estimated in simulators.

[0069] The characteristics of electric machine 3 are usually Diagnostic characteristics of non-standard machine operation, life time, incipient fault initiation characteristics (faults), or machine operating characteristics. The onset characteristics of incipient failures, generally faults in the electrical machine, do not at first impair its operation, but alter the electromagnetic field generated by the machine 3. These faults, if not identified, and extinguished lead to electrical machine to a fault. Accordingly, in a preferred embodiment of the invention, the supervisory system 6 shuts down the electric machine 3 if the characteristic is a fault characteristic.

[0070] The operating or operating diagnostics characteristics may be, among others, a fault condition, a fault condition, a normal operating condition, the life of the machine, among others. The fault condition of the electric machine 3 can be, among others, a damper winding malfunction, field winding short circuits, bus faults, voltage unbalance, phase unbalance, loop short circuits, short circuits. phase-to-circuit, open coils, open electrical circuits, eccentricities, mechanical vibration, iron shift, external short circuits, loss of power to machine parts, excitation problems, mechanical bearing or mechanical balance problems, loss of insulation, short circuit in magnetic cores, other problems in stator and rotor magnetic cores, leakage currents, improperly induced currents, undervoltage, overvoltage, presence of harmonics in the supply, overheating, influences of electrical or mechanical load, among others. Normally, when a failure or fault occurs in the electrical machine, the frequency spectrum of the external magnetic field (or magnetic induction, or magnetic flux) changes, especially in the low order and even order frequency components, facilitating thus their identification by the system and method of the invention. However, the system and method of the invention also monitor high frequency components up to the order of magnitude ten thousand times the frequency of the electric fundamental.

In Figure 2, a system diagram according to another preferred embodiment is shown, wherein the time derivative of an electromagnetic quantity 501 generated by the electric machine 3 is the time derivative of the grounding current of the electric machine. In this embodiment of the invention the method for identifying characteristics of an electric machine 3 according to the invention is also applied.

The system of figure 2 comprises at least one non-invasive meter 101 positioned externally to the electric machine 3 in at least one position 201 with a convenient orientation. In this embodiment of the invention, the meters may be unidirectional, three-dimensional or point depending on the type of sensor used. In a preferred embodiment, a Rogowisky coil is used as a meter, which naturally provides the waveform values proportional to the first order derivative of the grounding current of the electric machine 3.

The position 201 shown on the ordinate axis below the electrical machine of figure 2 corresponds to the distance of meter 101 from the grounding cable of machine 3. The meter 101 may be located at a distance that may vary radially from the machine 3 or its grounding cable. This distance ranges from a position against the machine frame or grounding cable surface to a position located at a distance of 20 times the diameter of the machine 3. In the preferred embodiment of the invention wherein meter 101 is a coil of Rogowsky, this meter 101 is disposed around the grounding cable of the electric machine 3. As in the embodiment of FIG. 1, electric machine 3 has an electromagnetic signature corresponding to the grounding cable current. In a preferred embodiment, the time derivatives of the grounding wire current derivatives and / or the frequency derivative waveform components of the time derivatives of that current are analyzed. Different electrical machines or states of the same electrical machine comprise electromagnetic signatures of the grounding current and its distinct time derivatives. The electromagnetic signature in the embodiment of FIG. 2 is the frequency spectrum of the time derivatives of the grounding current 501 of the electrical machine 3 identified by meter 101.

[0074] In a preferred embodiment of the invention, the electrical machine grounding cable current signature 3 is comprised of the sub-synchronous, harmonic, and interharmonic components of the fundamental electric frequency, including the zero frequency component, or the frequency component. harmonics of the mechanical fundamental frequency, including the fundamental of mechanical rotation.

Whenever a problem occurs in electrical machine 3, the electromagnetic signature of the current and its time derivative values from the machine grounding cable 3 (including the waveform frequency spectrum components of these time derivatives) suffers a Amendment The system according to this embodiment can identify this change in the electromagnetic signature of machine 3 and relate this change to a characteristic of the electric machine.

At least one meter 101 identifies the values relating to the time derivative of the grounding current of the electrical machine 3. These values 501 may be values derived from order n, such that n may range from 1 to infinity, and is preferably 1.

These values 501 are obtained analogously by meter 101. Once obtained, the time derivative values of the electromagnetic quantity 501 are sent to the analysis device 4 which will perform the same signal treatment and identification characteristics of the electrical machine. 3 described for the embodiment of the invention of FIG. 1 and then will provide the identified features for display on an analysis device display 4, or for another form of analysis result presentation, such as, for example, a listing of values, or will make the data available to a supervisory system 6 included or not in the analysis device 4.

In Figure 3, a block diagram of the method steps for identifying characteristics of an electric machine according to the invention is shown. In step 1000, the time derivative values of an electromagnetic quantity 50, 51, 52, 53, ... 5N, 501 generated by the electric machine 3 in at least one position 2, 22, ..., 2N are identified. , 201 external to the electric machine 3. This step 1000 can be performed by means of a meter 1, 1 1, ..., 1 N, 101 that measures the scalar or vector values of the electromagnetic quantity50, 51, 52, 53 , ... 5N, 501. In addition, derivative operators may be applied, for example, to scalar or vector values, or data processing steps may be performed to identify the values of the temporal derivative of the quantity. 50, 51, 52, 53, ... 5N measured and / or of frequency spectrum waveform components of the time derivatives of these quantities.

In step 1001 the comparison 1001 of the identified values of the time derivative of the electromagnetic quantity 50 is performed, 51, 52, 53, ... 5N, 501 generated by the electrical machine 3 and / or of the time derivative waveform frequency spectrum components of these quantities with the pre-established time derivative values and components of their time spectrum. frequency of that electromagnetic quantity generated by the electric machine 3.

In step 1002, at least one characteristic of electric machine 3 is determined 1002 based on the result of the comparison. These method steps can be performed by the system to identify characteristics of an electrical machine 3 of the type described herein.

In Figure 4, a block diagram of the steps of the method according to the invention is shown in a preferred embodiment, wherein the step of comparing 1001 the identified values of the time derivative of the electromagnetic quantity 50, 51, 52, 53 , ... 5N, 501 generated by the electric machine 3 with the preset values of the time derivative of that electromagnetic quantity further comprises the substeps of:

[0082] - filtering the measured signal 1003;

[0083] - amplifying 1004 said signal;

Passing this signal 1005 through an anti-aliasing filter 43; and

[0085] - discretize the signal 1006.

In substep 1003, the signal sent by meter (s) 1, 11, ..., 1 N, 101 is filtered through the filters 41 in the analysis device 4 of the system according to the invention. The filtered signal is then sent to be amplified at substep 1004 through instrumentation amplifiers 42, then at substep 1005 to pass an anti-aliasing filter 43. Then, at substep 1006, the analyzer 4 discretizes the signal through the sampler circuit 44, and finally a computational medium 45 reads the discretized signal and compares it to the pre-established values of the temporal derivative of that large one. 50, 51, 52, 53, ... 5N generated by the electric machine 3. In this way, the comparison step 1001 is completed, and then step 1002 of determining at least one characteristic of the electric machine 3 can be performed. .

[0087] In Figure 5, a block diagram of the steps of the method according to the invention is shown in another embodiment of the invention, wherein, upon determination 1002 of at least one feature of the electric machine 3, the collected data is 1007 to be used as preset values of time derivative of that electromagnetic quantity generated by machine 3. These data stored as preset values are used to obtain a first electromagnetic signature of the electric machine 3. These stored values can also be used to constitute a history of operation of the electric machine 3 comprising various values identified with a given time interval.

In Fig. 6 is shown a frequency spectrum from the first order time derivative of the electromagnetic magnitude values. In this embodiment of the invention, the electromagnetic magnitude is the magnetic field external to the machine 3. Three meters are used, one fixed to the outer wall and the other to the inner wall of the housing compartment of a generator of a hydroelectric plant and the third meter to the accessible face of the machine's stator housing. In Figure 6, the meters are identified as: meter 1 (housing external face), meter 2 (housing internal face) and meter 3 (stator housing face). The electric machine 3 corresponding to this frequency spectrum is a synchronous generator operating at 60Hz without fault. Note the presence of subharmonics (frequency components smaller than the 60Hz electrical fundamental frequency) and interharmonics (distinctive frequencies). 60Hz integer multiples) naturally present in the machine due to small electromagnetic or constructive machine asymmetries. The mechanical fundamental frequency is 2Hz, the other components being their natural multiples. By the method and system of the present invention all of these components are identified, their signal is treated, and compared to pre-set values for the external magnetic field 3 of the machine and its temporal derivative (s). to determine at least one characteristic of the electric machine 3.

Since the amplitudes of the subharmonic and interharmonic components, or the mechanical fundamental frequency and their harmonics, are relatively smaller than the electrical fundamental frequency, the analysis of the amplitudes of the frequency components of interest applied in the invention is performed. in decibel and in values relative to a frequency which may be electrical fundamental, or mechanical fundamental frequency or any other, as long as it is suitable for the spectral analysis procedure and the machine characteristic analysis 3. This is one of the advantages of the invention because subharmonics and interharmonics (or the mechanical fundamental frequency and its harmonics) typically have amplitudes smaller than the electrical fundamental frequency amplitude that would, visually, make it impossible to detect anomalies in the machine if actual absolute values of quantities were used. as they would be at the beginning of the scale.

In Figure 7 a comparative example of the magnetic signatures of the spectra of two generators is shown. In this plant, the evaluated generators A and B have the same construction design, so that a similar spectral pattern was found between them. The small amplitude differences observed for some components may indicate particularities of each equipment. Also in these experiments, the sensors were positioned at the same point in relation to each generator's statistic package, which worked under the same operating condition.

[0091] Figure 8 shows a comparative example of the electromagnetic signature of two synchronous generators operating at 60 Hz, showing the spectral analysis of the first order time derivative values of the generators grounding cable current (neutral current) as electrical quantity. These currents show the presence of some of the subharmonic frequency components. Moreover, comparing, for example, the spectrum of these currents for generators A and B of Figure 7, it is noted that this quantity has its own signature for each electric machine.

[0092] Figure 9 shows an example of identifying the amplitude variation of the subharmonic components as a function of the severity of the short circuit type field winding of an eight-pole synchronous generator operating at 60Hz. The 15 Hz frequency component is the fundamental mechanic. In this figure, the amplitude values in dBV of the subharmonic components of the first order derivative of the external magnetic field relative to the electrical fundamental frequency measured by a system meter according to the invention, initially the machine, are shown. is short-circuited, and subsequently with a progressive increase in the number of turns shorted.

To exemplify the detection of anomalies in an electrical machine, as can be seen in Figure 9, as the fault increases (in this case, the increase in the number of shorted turns), the amplitudes of the subharmonic components. monitored increase. When changing the sound state to 20% turns of a pole shorted, all three components observed subharmonic levels increase in amplitude, in slightly different proportions, but by approximately 12 dBV.

[0094] As the fault increases, the subharmonic amplitudes continue to grow, again in distinct proportions. For a short circuit in 50% of the shorted turns, there was an average increase of 7.5 dBV from the previous state, or 19.5 dB from the considered healthy state. Shorting all the turns of a pole, the amplitudes increased by approximately 5dBV for the subharmonic components at 15 and 30 Hz, and 2.5 dBV for the 45 Hz component. The amplitude of the mechanical fundamental frequency of 15 Hz was about the same amplitude as the electric fundamental frequency. The measurement, analysis and evaluation of the behavior of the fundamental mechanics and their harmonics present in the external magnetic field time derivative waveforms (or external magnetic induction or flux) is also a novelty of the invention. In this modality, only one sensor was employed, proving sufficient to detect the incipient fault.

Through these differences in amplitudes when the electrical machine goes missing, fails, or when something is wrong with the machine, the system and method can identify by measuring the time derivative values of an electromagnetic quantity. , 51, 52, 53, ... 5N, 501 generated by electric machine 3 in the frequency domain, at least one characteristic of electric machine 3.

The system and method for identifying features of an electric machine according to the invention therefore non-invasively perform the identification of time derivative values of an electromagnetic quantity 50, 51, 52, 53, .. .5N, 501 and / or time derivative waveform frequency spectrum components of this magnitude generated by the electrical machine 3 in at least one position 2, 22, 2N, 201 which may vary radially with respect to machine 3 from abutting the machine casing to a distance 30 times the diameter of machine 3, to then compare the identified values of the time derivative of the machine. electromagnetic quantity generated by electric machine 3 with the pre-established time derivative values of that same electromagnetic quantity. Then, in the analysis device 4 at least one characteristic of the electric machine 3 is determined based on the result of the comparison.

All steps of the method for identifying characteristics of an electrical machine described herein may be performed by means of the system for identifying characteristics of an electrical machine described herein.

Having described a preferred embodiment example, it should be understood that the scope of the present invention encompasses other possible variations, being limited only by the content of the appended claims, including the possible equivalents thereof.

Claims

1. system for identifying characteristics of an electrical machine (3) comprising at least one non-invasive meter (1, 1 1, ..., 1 N, 101) and an analysis device (4);

characterized in that the at least one meter (1, 1 1, ..., 1 N, 101) identifies time derivative values of an electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) generated by the electric machine (3) and sends these time derivative values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) to the analysis device (4), and

the analysis device (4) interprets the time derivative values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) identified by at least one meter (1, 1 1, ..., 1 N, 101) noninvasive and determines at least one characteristic of the electric machine (3).

A system according to claim 1, characterized in that the electromagnetic magnitude is at least one component of the waveform frequency spectrum and its temporal derivatives of one of the magnetic field external to the electric machine (3); the magnetic flux external to the electric machine, the magnetic induction external to the electric machine and the grounding current of the electric machine.

System according to claim 2, characterized in that each meter (1,11, ..., 1 N, 101) measures at least one different component of the frequency spectrum of the electric machine.

System according to one of Claims 1 to 3, characterized in that the analysis device (4) compares the identified time derivative values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) with predefined time derivative values of this electromagnetic quantity and determines at least a feature of the electric machine based on the comparison result.

System according to any one of claims 1 to 4, characterized in that the time derivative values of the identified electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) correspond to the time derivative of n order of said electromagnetic quantity, wherein n ranges from 1 to infinity.

System according to any one of claims 1 to 5, characterized in that the at least one non-invasive meter (1, 1 1, ..., 1 N, 101) is at least one of: a meter punctual, unidirectional and three-dimensional.

System according to any one of claims 1 to 6, characterized in that each at least one non-invasive meter (1, 11, ..., 1 N, 101) comprises at least one of:

an analog sensor identifying the time derivative values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) and

a sensor comprising a calculation means, wherein the sensor measures scaling or vectorically the values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501), and the calculation means applies derivative operators to the values measured by the sensor, and identifies the time derivative values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501).

System according to any one of claims 1 to 7, characterized in that each at least one non-invasive meter (1, 1, ..., 1 N, 101) is arranged in a position (2, 22, ..., 2N) between a location in contact with the machine housing and a distance away from the electric machine (3) of up to 30 times the diameter of the electric machine (3).

System according to any one of claims 1 to 8, characterized in that each at least one non-invasive meter (1, 1, ..., 1 N, 101) is arranged in a position (2, 22, ..., 2N) surrounding the grounding cable of the electric machine (3).

A system according to any one of claims 1 to 9, characterized in that the identified electrical machine characteristic is at least one of: non-standard machine operating diagnostic characteristics, lifetime, incipient fault initiation, fault status, or machine operating characteristics.

1. Method for identifying characteristics of an electrical machine (3), characterized in that it comprises the following steps:

- identify (1000) time derivative values of at least one electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) generated by the electric machine (3) in at least one position (2, 22, .. ., 2N, 201) external to the electric machine (3);

- compare (1001) the identified values of the time derivative of at least one electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) with the pre-established values of the time derivative of this electromagnetic quantity generated by the electric machine ( 3); and

- determine (1002) at least one characteristic of the electric machine (3) based on the result of the comparison.

Method according to claim 11, characterized in that the electromagnetic quantity is at least one component of the frequency spectrum of one of the magnetic field external to the electric machine (3), the magnetic flux external to the electric machine , the external magnetic induction to the electric machine and the grounding current of the electric machine.

The method according to claim 11 or 12, characterized in that the step of identifying (1000) the value of the derivative The time period of the electromagnetic magnitude is performed by at least one non-invasive meter (1, 1 1, ..., 1 N, 101) from:

an analog sensor for identifying time derivative values of an electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) and

a sensor comprising a means of calculation, wherein the sensor measures the values of the electromagnetic quantity (50, 51, 52, 53, ... 5N, 501), and the means of applying derivative operators to the values measured by the sensor, and numerically identifies the time derivative values of this electromagnetic quantity (50, 51, 52, 53, ... 5N, 501).

Method according to claim 13, characterized in that each meter (1, 11, ..., 1 N, 101) identifies a different component of the frequency spectrum of the electric machine.

Method according to any one of claims 11 to 14, characterized in that the time derivative values of the at least one identified electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) correspond to to the time derivative of order n of said electromagnetic quantity, where n ranges from 1 to infinity.

Method according to any one of claims 1 to 15, characterized in that the at least one non-invasive meter (1, 11, ..., 1 N, 101) is one of a point meter, uni - directional or three-dimensional.

Method according to any one of claims 11 to 16, characterized in that at least one position (2, 22, ..., 2N, 201) external to the electric machine (3) where the The step of identifying the time derivative values of at least one electromagnetic quantity is between a location in contact with the machine frame and a distance apart. 30 times the diameter of the electric machine (3).

A method according to any one of claims 11 to 17, characterized in that the step of identifying

(1000) Time derivative values of at least one electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) are performed around the grounding wire of the electrical machine (3) at a location between the grounding cable surface and a distance of a maximum of 30 times the diameter of the electric machine (3).

A method according to any one of claims 11 to 18, characterized in that the identified electrical machine characteristic is at least one of: non-standard machine operating diagnostic characteristics, lifetime, characteristics incipient failures, fault status, or operating characteristics of the machine.

A method according to any one of claims 11 to 19, characterized in that the step of comparing

(1001) Identified values of the time derivative of an electromagnetic quantity (50, 51, 52, 53, ... 5N, 501) generated by the electric machine (3) with the pre-established values of the time derivative of this electromagnetic quantity (50). , 51, 52, 53, ... 5N, 501) further comprises the substeps of:

- filtering (1003) the measured signal;

- amplifying (1004) the signal;

passing (1005) the signal through an anti-aliasing filter (43); and

- sampling (1006) the signal.

Method according to any one of claims 11 to 20, characterized in that it further comprises a step of storing (1007) the identified time derivative values of the electromagnetic quantity (50, 51, 52, 53, ..). .5N, 501) generated by the electric machine (3) to be used as pre-established time derivative values of the electromagnetic quantity.

Method according to any one of claims 11 to 21, characterized in that the predetermined time derivative values of the electromagnetic quantity correspond to the operating measurement values of the sound electric machine (3), historical values. of the electric machine (3) or expected theoretical values.

Method according to any one of claims 11 to 22, characterized in that the electromagnetic quantities are identified in dB.

Method according to any one of claims 11 to 23, characterized in that it is carried out by a system as defined in any one of claims 1 to 10.